aws-kinesis-writable

Kinesis writteable stream for bunyan.

Installation

npm install aws-kinesis-writable --save

Usage

var KinesisWritable = require('aws-kinesis-writable');

var kinesis = new KinesisWritable({
  accessKeyId:     'KEY_ID',
  secretAccessKey: 'SECRET_KEY',
  region:          'AWS_REGION',
  streamName:      'MyKinesisStream',
  partitionKey:    'MyApp'
});

process.stdin.resume();
process.stdin.pipe(kinesis);

Configuration Parameters

buffer (defaults to true): It can be a boolean or an object describing its conditions.

This library uses by default an smart buffering approach. Messages are sent when one of the following conditions are meet:

• X seconds after the last batch of messages sent. Default: 5 seconds.
• X messages are queued waiting to be sent. Default: 10 messages.
• a message is prioritary. Default: all messages are not prioritary

Example:

new KinesisWritable({
  region:          'AWS_REGION',
  streamName:      'MyKinesisStream',
  partitionKey:     "foo",
  buffer: {
    timeout: 1,                         // Messages will be send every second
    lenght: 100,                        // or when 100 messages are in the queue
    isPrioritaryMsg: function (msg) {   // or the message has a type > 40
      var entry = JSON.parse(msg);
      return entry.type > 40;
    }
  }
});

partitionKey can be either an string or a function that accepts a mesage and returns a string. By default it is a function that returns the current EPOCH (Date.now()). Example:

new KinesisWritable({
  region:          'AWS_REGION',
  streamName:      'MyKinesisStream',
  partitionKey:     function (msg) {
                      var entry = JSON.parse(msg);
                      return entry.level + '|' + entry.name;
                    }
});

streamName is the name of the Kinesis Stream.

Methods

getStreamName(): returns Stream's name.

setStreamName(name): set the name of the stream where messages will be send.

Events

errorRecord: Emitted once for each failed record at the aws.kinesis.putRecords's response.

Note: Amazon Credentials are not required. It will either use the environment variables, ~/.aws/credentials or roles as every other aws sdk.

License

MIT 2015 - AUTH0 INC.